Vol. 2, NO.4,July-Aug., 1963
I.
SPECIFICITY OF
B. subtih RIBONUCLEASE
787
Studies on B. subtilis Ribonuclease. Characterization of Enzymatic Specificity.
G. W. RUSHIZKY, A. E. GRECO,R. W. HARTLEY, Jr.,
AND
H. A. SOBER
From the Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Public Health Service, U. S. Department of Health, Education and Welfare, Bethesda 14,Maryland Received March 6, 1963 B. subtilis ribonuclease was extensively purified from a bacterial medium with a yield of 11%. The enzyme was found to be free of contaminating ribonucleases and phosphatases. The course of enzymatic action was examined by characterization of different stages of digestion of high molecular weight RNA and various oligonucleotides. B. subtilis ribonuclease was thus found to hydrolyze -GpGpand -GpAplinkages about 100 times faster than other phosphodiester bonds, but did not have a sole specificity for purines as reported previously. At the last stage of digestion, only mono- and dinucleotides remained. Possible applications of the preferential specificity of the enzyme to the characterization of nucleotide sequences and preparation of large oligonucleotides are discussed.
B. subtilis produces an extracellular ribonuclease (Nishimura and Nomura, 1958, 1959) which was reported to hydrolyze only the secondary phosphate esters of the purine ribonucleoside 3 '-phosphates of yeast ribonucleic acid (RNA). Thus a t the end of digestion the mononucleotide fraction and the end groups of the isolated oligonucleotides consisted only of guanylic and adenylic acids. Among the mononucleotides produced, the ratio of guanylic to adenylic acid was about 3:l (Nishimura, 1960). By comparison, pancreatic ribonuclease which is specific for the hydrolysis of phosphodiester bonds between pyrimidine riboside 3'-phosphates and adjacent purine or pyrimidine nucleotides (Schmidt et al., 1951) releases cytidylic and uridylic acids (in yeast RNA) in a ratio of about 1:l (Markham and Smith, 1952; Volkin and Cohn, 1953). Since a purine-specific enzyme would be very useful for the determination of nucleotide sequences in RNA, it was of interest to characterize further the specificity of B. subtilis ribonuclease as well as to explain the reason for the unexpected ratio of purine mononucleotides released in B. subtilis ribonuclease digests of RNA. In this report we describe the extensive purification of B. subtilis ribonuclease. The second paper of this series (Hartley et al., 1963) deals with the physical properties of this purified enzyme. Using such enzyme preparations, digests of high molecular weight RNA as well as of various oligonucleotides of known base-sequences were examined. Contrary to the previously reported results of Nishimura and Nomura, our data show that while B. subtilis ribonuclease preferentially hydrolyzes -Gp/Ap- and -Gp/Gp-l linkages, other internucleotide bonds in RNA are hydrolyzed as well. Only mono- and dinucleotides are found in a complete digest. MATERIALS AND METHODS
All spectrophotometric measurements were made with a Beckman DU spectrophotometer with silica cells of 1-cm light path and are expressed as absorbancy (A). A Leeds and Northrup meter equipped with microelectrodes was used for pH determinations. For paper chromatography with Whatman No. 3MM paper, two solvents were used, Solvent A (n-propanolconcentrated NH40H-H20,55:10:35 by volume) and 1 The abbreviations are those in use by the Journal of Biological Chemistry. A slanted Line denotes hydrolysis of the internucleotide bond.
Solvent B (40 g of ammonium sulfate added to 100 m of 0.1 M sodium phosphate, p H 7.0). Alkaline phosphatase* from Escherichia coli was used for removal of the terminal 3'-phosphate (Khorana and Vizsolyi, 1961). Oligonucleotideswere hydrolyzed with snake venom diesterases that contained insignificant amounts of 5'-nucleotidase (Koerner and Sinsheimer, 1957). Pancreatic ribonuclease' was also used for this purpose under the same conditions as described previously (Rushizky et al., 1961). RNA.-High molecular weight RNA was prepared from pressed cakes of bakers' yeast by treatment with detergent (Crestfield et al., 1955). The fraction insoluble in M NaCl was then extracted with phenol. A 0.1% solution of this RNA in water a t neutral pH was calculated to have an A260 of 24.3. DNA.-Deoxyribonucleic acid from sperm was obtained from Nutritional Biochemical Corporation, Cleveland, Ohio. Oligonucleotides.-Compounds terminating in guanosine 3'-phosphate were obtained from ribonuclease-TI digests (Rushizky and Sober, 1962) of yeast RNA. The other oligonucleotides were isolated from pancreatic ribonuclease digests of the same RNA (Rushizky and Knight, 1960). Adsorbents.-DEAEand CM-cellulose (Peterson and Sober, 1956) (0.8 and 0.6 meq/g, respectively) were obtained from Brown Co., Berlin, N. H. Before use, fines were removed by repeated decantation, and the adsorbent was sieved and washed (Peterson and Sober, 1956). Columns were packed with increasing pressure to 15 psi using 100-230 mesh adsorbent equilibrated with the starting buffer (Table I) and washed until the conductivity and pH of the effluent solution were the same as those of the influent solution. Assay for B. subtilis Ribonuclease.-To 0.8 ml of the assay solution (2.0 mg of yeast RNA per ml, and 0.125 M in Tris-C1, pH 8.5) was added 0.2 ml of properly diluted enzyme solution; the mixture was held a t 37" for 15 minutes. The reaction was stopped by the addition of 1 ml of 6% HC1O4. After 15 minutes a t O o , the precipitate was centrifuged out, 0.5 ml of the
* We are grateful to Father Donald J. Plocke, Peter Bent Brigham Hospital, Boston, Massachusetts, for a gift of this enzyme preparation. a Obtained from Crotalus adamanteus venom from Ross Allen's Reptile Institute, Silver Springs, Florida. 4 Pancreatic ribonuclease (twice recrystallized) was obtained from Worthington Biochemical Corporation, Freehold, N. J.
788
Biochemistry
G . W. RUSHIZKY, A. E. GRECO, R. W. HARTLEY, AND H. A. SOBER
TABLE I CHROMATOGRAPHY OF B. subtilis RIBONUCLEASE SOLUTIONS
Cellulose Adsorbents, (g)
DE-,
100 g C M ( I ) , 95 g
Column Dimensions (inside diameter X height in cm) 4.4 4.4
x 55 x 55
Starting Buffer
Limit Buffers.
0.01 MAmB,pH 8 . 6 0.001 M AmAc, p H 4 . 3 5
0.05~AmB,pH8.6 (A) 0 . 1 M AmAc, p H 4 . 3 5 (B) 0 . 1 M AmB, pH 7 . 6
CM(II), 30 g 2 . 2 X 75
0.005 M AmAc, pH 4.35
(A) 0.05 M AmAc,pH 4.35 (B) 0 . 2 M AmAc, p H 5 . 7
CM(III), 30 g 2 . 2
x
75
0.005~NaAc,pH4.4
(A) O . l ~ N a A c , p H 4 . 7 (B) 0 . 2 ~ N a A c , p H 5 . 7 0 . 1 M NaCl
Fraction of Limit Buffer* invarigrad Chambers 3, 4 5 6 7 8,9 2 3 4 5,6 7, 8 9 2 3 4, 5 6, 7 8,9
=
= = =
= =
= = =
=
= =
= = =
=
0.5A A 0.1B 0.5B B 0.2A A 0.05B 0.25B 0.5 B B 0.5 A A 0.25 B 0.5 B B
Gradient Flow Vol- R a t e ume (ml/ (ml)
2700
hr) 100 36
2250
32
2250
32
If not indicated in this column, the chambers of the Varigrad (Peterson and Rowland, 1961) contain starting buffers only. For example, in C M ( I ) , chambers 1 and 2 contain the starting buffer only. Abbreviations used for designation of buffers: AmB, ammonium bicarbonate, AmAc, ammonium acetate, NaAc, sodium acetate. b Peterson and Sober (1959).
supernatant diluted with 4.5 ml of water, and the AZM)determined against the proper blank. This assay was linear up to Az, values of about 0.7; there was no detectable lag period. An increase in Az, of 1.0 under these conditions was defined as 100 units of enzyme activity. Purification of B. subtilis ribonuclease.-B. subtilis strain H (Nishimura and Nomura, 1958) was used to inoculate 150 ml of a medium containing 0.5% lactose, 1%neopeptone, 0.5% meat extract, 0.2% yeast extract, and 0.2% NaC1, a t p H 7.2.5 The inoculum was grown for 24 hours a t 30' with aeration by shaking, assayed for B. subtilis ribonuclease activity (100-500 unita per ml), and examined under the microscope for the charadenstic rods of B. subtilis. Portions of 25 ml of inoculum were then transferred to each of six 4-liter Erlenmeyer flasks containing 1 liter of Nishimura's medium (Nishimura and Nomura, 1958) modified to contain 2% neopeptone in place of soybean extract. After shaking for 24 hours a t 30°, the maximum amount of B. subtilis ribonuclease was obtained. The 6 liters of medium were pooled, the cells centrifuged out, and the supernatant solution (at 0 " ) brought to 0.4 M Hi% (PH 3.68 determined a t a 1:lOOO dilution) with concentrated HzS04. After 16 hours a t 5", the p H of the solution was adjusted to 2.3 (undiluted) with conc. ",OH; 500 g of (NH&SOa per liter of solution was then added. Due to the high density of this solution, the enzyme and other protein rose to the top upon standing for 24 hours a t 5". The clear yellow solution below the protein layer was drawn off; the residue was suspended in 1 liter of water and adjusted to p H 7 with NH40H. After extensive dialysis a t 4" against three 10-liter volumes of water for 12-16 hours each, insoluble material was centrifuged out and the supernatant solution was then applied to a DEAE-cellulose column (Table I). The enzyme (at p H 8.6) passed through the column without being bound to the adsorbent; 5 We are grateful to Drs. S. Nishimura and H . Tizuka of the Institute of Applied Microbiology, University of Tokyo, for a gift of this culture.
the last traces of B. subtilis ribonuclease activity were eluted with 0.05 M NH4HC03,p H 8.6. The crude B. subtilis ribonuclease solution (about 2 liters) was adjusted to p H 4.3 with 1 ml of glacial acetic acid and adsorbed on a CM-cellulose column previously equilibrated with 0.001 M ammonium acetate, pH 4.35. (CM [I], Table I). Two peaks with ribonuclease activity were thus isolated. Peak A was not bound to the adsorbent and passed through the column together with all the pigment which had accompanied the B. subtilis ribonuclease up to this point. Peak B was eluted after 81% of the total gradient volume had gone through the column. The fractions containing the highest specific activity were pooled, dialyzed, adjusted (by dilution) to the concentration of the starting buffer of the next column (no pH adjustments were necessary), and further purified by two chromatography steps on CM-cellulose (Table 11). In both buffer systems the enzyme was eluted after 75-90% of the total gradient volume had passed through the column. After the third chromatography on CM-cellulose, the peak fractions were pooled and stored a t 3" in the presence of chloroform a t a concentration of about 1 mg/ml. B. subtilis ribonuclease does not lose activity under these conditions for several months. Characterization of Ribmuclease Activity in Peak A.An aliquot of about one-tenth of the pooled material in peak A was dialyzed and resubmitted to chromatography as described for CM-cellulose I (Table I). Other aliquots of peak A were compared with further purified B . subtilis ribonuclease from peak B with respect to p H optimum, effect of EDTA, CuS04, MgS04 (Table III), and enzymatic action on yeast RNA as described below. Assay of B. subtilis Ribonuclease for Contaminating Enzymes.-Purified B. subtilis ribonuclease was tested for acid phosphatase and nonspecific phosphodiesterase (Koerner and Sinsheimer, 1957) as well as for deoxyribonuclease activity (Rushizky and Sober, 1962). Mapping and Digestion of Yeast RNA and Oligonucleotides with B. subtilis Ribonuclease.-The mapping procedure has been described (Rushizky and Knight,
Vol. 2, NO.4, July-Aug., 1963
SPECIFICITY OF
TABLE I1 PURIFICATION OF B. subtilis RIBONUCLEASE FROM 6 OF BACTERIAL MEDIUM
LITERS
460,000 375,000 91,000 44,600 12,000
49 56 52 37 33
106 150 574 828 2,740
(11)
12
After C M -cellulose chromatography
10 17 13
5.4
789
TABLEI11 EFFECT OF METAMAND E D T A ON ACTIVITYOF B. subtilis Compound Tested EDTA
100 114 106 75 66
cuso, MgSO4
10,000 1,000 122
RIBONUCLEAEE
RIBONUCLEASE"
Total Enzyme % Protein Units Specific Re(as (X * Activity* covery Original medium After cell removal WHdzSOi Ppt After dialysls After DEAE-cellulose chromatography After CM-cellulose chromatography (1) Peak AC Peak B After CM-cellulose chromatography
B. SUbtdliS
912 16,100 105,300
20 35 26
446,000
11
(111) F o r a description of enzyme units and enzyme assay, see text. * Total units of enzyme/total A280. For a characterization of the enzymatic activity in these peaks, see text. a
1960). To check for the completion of digestion, equal amounts of RNA in 0.1 M Tris-C1, p H 8.5, were hydrolyzed a t increasing enzyme-substrate ratios (see legend to Table IY). Digestion was halted by treatment with phenol followed by extraction with ether. Recovery of nucleic acid material has been shown to be quantitative under these conditions (Rushizky et al., 1961). A separate portion of the enzymatic [digest was hydrolyzed with N KOH (24 hours a t 23') and assayed spectrophotometrically in order to measure the total amount of nucleotide material used for mapping. The mono- and oligonucleotides eluted from maps of the RNA digests were identified as previously described for the same compounds obtained after digestion with micrococcal nuclease (Rushizky et al., 1962) and ribonuclease-T1 (Rushizky and Sober, 1962). Compounds terminating in 2',3'cyclic-terminal phosphates were separated from their
Per Cent Activity Peak A Peak B
M
0.001 0.01 0.1 0.001 0.01 0.1 0.001 0.01 0.1
- 3
- 11
+ 5
+ 5
- 1 - 14 - 98 - 16 - 37 - 83
- 8 - 3 - 41 - 96 - 28 - 38 - 81
- 13
a F o r a definition of peaks A and B, see Table I1 a n d t h e text.
corresponding 3'-phosphate forms by paper chromatography with solvent B. Oligonucleotide digests were also freed of enzyme with phenol, and fractionated by paper chromatography with solvent A, B, or by mapping. Determination of the nucleotide composition of the oligonucleotide produds in the hydrolyzates was facilitated by eluting the spots from the chromatograms for 3 hours with 0.01 M (0.1 M for papers developed with solvent B) ammonium acetate, p H 4.5, containing 0.02 units of ribonuclease T2per ml. Under these conditions oligonucleotides are completely hydrolyzed to nucleoside 3 'phosphates, which can be quantitatively determined by spectrophotometry a t p H 7.0 (Rushizky and Sober, 1963). RESULTS
B . subtilis ribonuclease was purified from 6 liters of bacterial medium by treatment with acid and ammonium sulfate, and by chromatography on DEAEand CM-cellulose (Table I). The total purification was 4200-fold, and the yield 11% (Table 11). The p H optimum in 0.1 M Tris-C1 was a t p H 8.5, with 50% of maximal activity remaining a t p H 7.5 and 9.3. At p H 8.5, varying the substrate concentrations from 0.75 to 7.5 mg of RNA per ml had no effect on enzyme activity. EDTA (0.001-0.1 M) also did not affect the enzyme; cdo4 and MgS04were inhibi-
TABLE IV MONO-A N D OLIGONUCLEOTIDES IDENTIFIED IN B. subtilis RIBONUCLEASE DIGESTS OF YEASTR N A
Staae 11' G-cyclic-p
Stage 2 Gp,G-cyclic-p APb ADAD
Stane 3
spot Number (Fin. 1) 4, 5
Gp,G-cyclic-p AP,CP UP CDAD.ADAD
10 2
UPUP (APUPUP),(UPUPCP) CPUPGP,UPCPGP,APUP~P APCPGP,CPAPGP,CPCPGP APUPAP, (CPUPAP),UPCPCP APAPCP,APCPCP (APAPUPUP),(CPCPUP)GP"
13 14 12 7 9 3 15
1
Stage 4 Gp,G-cyclic-p AP7CP UP
8
(CPUP)GP, (APUP)GP (APCP)GP,APAPGP,CPCPC.P (APCPUP)GP, (CPCPUP)GP
(CPUPIGP CPCPGP APCPCP
a Digestion conditions: 3.5 mg of R N A in 0.5 ml of 0.1 M Tris-C1, p H 8.5, were digested for 6 hours at 37O with 1.75, 195, and 3850 units of B. subtilis ribonuclease added in 0.1 ml of water. This corresponds t o stages 1-3, respectively. Stage 4: 3.5 mg of R N A in 0.4 ml of buffer as above were digested for 8 hours at 37" with four equal portions (total units: 77,000) of enzyme added at 2-hour intervals. For a definition of enzyme units, see text. * Except for t h e case of Gp, compounds terminating in 2',3'-cyclic-terminal phosphate and 3'-phosphate are listed together. c Stages 1-2 contained core material, while in Stage 3 only tetra- and in Stage 4 only tri- and smaller oligonucleotides were found.
790
Biochemistry
G. W. RUSHIZKY, A. E. GRECO, R. W. HARTLEY, AND H. A. SOBER
TABLE V HYDROLYSIS OF TRINUCLEOTIDES BY B . subtilis RIBONUCLEASE^
Compound
Units of Enzyme per mpmole of Compound
(CPUP)GP
2.9
(CPUP)GP % hydrolysis
APUPGP
1.9
Ap,A-cyclic-p UPGP APUPGP
mpMoles of Compounds Isolated after Varying Times of Digestion (hours) at 37'
Compounds Isolated
Deduced
0
740
841 AP, UPGD
% hydrolysis APAPCP
1.6
Ap,A-cyclic-pL CP AbAP APCP APAPCP % hydrolysis
APAPUP
1.6
Ap,A-cyclic-p APUP UP APAP APAPUP %
APGPUP
1 75
900 APAP 'CP AP, APCP
hydrolysis
Ap,A-cyclic-p GPUP UP APGP APGPUP
460 APAPI UP AP/APUP
690 AP GPUP APGPI UP
% hydrolysis APGPCP
2.2
Ap, A-cyclic-p GPCP APGP CP APGPCP
590
ApGp,'Cp AP, GPCP % hydrolysis APGPCP
0.3
APGPCP % hydrolysis
1
2.9
743 0
24 25 820 25 3%
57 58 786 58 7%
87 50 51 91 745 51 89 16%
110 63 60 112 710 62 112 20 %
47 46 29 29 379 29 46 17%
69 65 30 29 357 29 68 21 %
25 23 90 89 570 24 90 16%
57 55 153 157 476 55 155 30 %
9 9 55 57 523 56 9 11%
17 16 78 79 490 79 17 16 %
75
Up,U-cyclic-p APGP UPAPGP
78 0
728
239 236 486 238 33 %
500 486 240 493 67 %
630
227 230 400 230 37%
447 440 191 440 70 %
UP, APGP 70 hydrolysis
APAPGP
2.0
Ap,A-cyclic-p APGP APAPGP
AP~APGP
70 hydrolysis
GPGPCP
0.05
G-cyclic-p GPCP GPGPCP % hvdrolvsis
GPAPCP
0.03
79 GP/GPCP
G-cyclic-p,Gp APCP GPAPCP
77 78 0 100% 78 80
80
0
Gp,'ApCp % hydrolysis
4
745 0
___
UPAPGP
2
100%
Val. 2, No. 4. July-Aug., 1963
SPECIFICITY OF
B. subtilis
791
RIBONUCLEASE
TABLE V (continued) mrMoles of Compounds Isolated after Varying Times of Digestion (hours) a t 37”
units of Compound
Enzyme per mpmole of Compound
Compounds Isolated ~
GPAPUP
0.01
Deduced ~
~
0
~~
G-CYC~~C-P APUP GPAPUP GP/APUP
493
% hydrolysis GPGPUP
0.01
G-cyclic-p GPUP GPGPUP
797 GP/GPUP
% hydrolysis
2
1 ~
~
4
~~~
193 199 304 196 40%
310 307 185 309 63 %
247 241 553 244 31%
For a description of the isolation procedure by paper chromatography and mapping, see text. T h e hydrolysis of a phosphodiester bond is indicated by a slanted line, /. * T h e progress of hydrolysis of the trinucleotides was deduced as follows: I n the case of ApApCp (1 hour digestion in 0.1 M Tris-C1, p H 8.5), 51 millimicromoles (mrmoles) of ApAp and 50 mrmoles of C p were found, accounting for 51 mrmoles of the original substrate hydrolyzed as ApAp/Cp. Similarly, 87 mpmoles of Ap plus A-cyclic-p were correlated with 91 mpmoles of ApCp, giving 89 mpmoles of Ap/ApCp. I n all cases, per cent hydrolysis is expressed in terms of the fraction of original trinucleotides split. a
tory, but only a t rather high concentrations (Table 111). The enzyme did not lose activity upon treatment with 0.4N H$Oa for 18 hours a t 4’. B. subtilis ribonuclease was inactive toward pnitrophenylphosphate and bis(p-nitropheny1)phosphate. When DNA was incubated for 8 hours a t 37” with 22,000 units of enzyme per mg substrate in 0.1 M Tris-C1, p H 8.5, no mono- or oligonucleotides were released as determined by paper chromatography using solvent B. After the first column chromatography on CMcellulose, two forms of B. subtilis ribonuclease were isolated. One of these, (peak A, Table 11) was not bound to the adsorbent a t p H 4.3 and had a considerably lower specific activity than peak B. Even small amounts of peak A material were not bound to CMcellulose under identical conditions. However, the two forms of B. subtilis ribonuclease did not differ from each other with respect to p H optimum, effect of EDTA, CuS04, MgS04, and enzymatic action on yeast RNA. When B. subtilis was grown as described but 0
Q @ n
FIG.1.-Drawing of a map showing the fractionation on Whatman No. 3 paper of 3.5 mg of R N A digested by B . subtilis ribonuclease (stage 3, Table IV). First dimension: electrophoresis (left t o right) in 0.02 M ammonium formate, pH 2.7,for 17 hours a t room temperature and 6 v ’cm. Second dimension: paper chromatography in descent, with 55:45 (v/v) tertiary butanol: 0.02 M ammonium formate, p H 3.8. T h e paper was serrated so t h a t t h e solvent could run off the paper without edge effects. Running time was 36-40 hours at room temperature. For a n identification of the compounds, see Table 4.
in a 300-liter batch with aeration by forced air rather than shaking, no peak A ribonuclease was obtained by the same purification procedure. The course of enzymatic action on RNA was followed by examining RNA digests prepared with 0.5, 55, 1100, and 22,000 units of B. subtilis ribonuclease per mg of substrate (stages 1 4 , Table IV). The h t monoand oligonucleotides released were G-cyclic-p and ApGcyclic-p (stage 1). At the next stage compounds terminating in Gp predominated with traces of Ap and ApAp present as well. Further enzymatic action (stage 3) resulted in the liberation of Up and Cp and significant amounts of oligonucleotides free of guanylic acid, in addition to the compounds already found in. stage 2. “Core” material (large oligonucleotides that, did not move during paper electrophoresis and chro matography) was evident until stage 3, but absent. from stage 4 digests. At this last stage, mono- and dinucleotides and only traces of trinucleotides were found. In order to compare our results with those of Nishimura (1960),the ratio of Gp to Ap (found in mononucleotide form) was determined. At stage 3 the ratio was 2.9, while a t stage 4 it was 1.6. The compounds listed in Table 4 were found in both the 2’,3’-cyclic-terminal phosphate and 3’phosphate form; the proportion of these was not o b tained for stages 1-3. In stage 4 the per cent (of the t,otal amount) of a compound isolated in the cyclicterminal phosphate form were as follows: Gp, 41: AP, CP, UP, 90-95; CPAP, APAP, UPAP, UPCP, and CpUp, 75-85; ApGp, CpGp, and UpGp, 5-10. These results are in line with the preferential hydrolysis of phosphodiester bonds between guanosine 3’phosphate and other bases observed in RNA (stages 1-2, Table IV). Thus, among the mononucleotides, the per cent of G-cyclic-p is lowest; similarly, dinucleotides terminating in Gp such as ApG-cyclic-p are hydrolyzed a t a faster rate than those without Gp, i.e., ApA-cyclic-p. The determination of enzymatic specificity of B . subtilis ribonuclease also involved the comparison of rates of hydrolysis of trinucleotides (Table V). (CpUp)Gp was hardly split under the conditions leading to partial cleavage of ApUpGp, UpApGp, ApGpUp, ApGpCp, ApApCp, and ApApUp. By comparison, the -GpGp- and -GpAp- sequence in GpGpCp, GpApCp, GpGpUp, and GpApUp was split about 1010 times faster than the six trinucleotides. The preferential hydrolysis of these two sequences is in agreement
792
0.W. RUSHIZKY, A. E.
GRECO, R. W. HARTLEY, AND
Biochemistry
H. A. SOBER
TABLE VI HYDROLYSIS OF DINUCLEOTIDES BY B . subtilis RIBONUCLEASE^
Compound
unite of Enzyme per m m o l e of Compound
mrMolea of Compounds Isolated after Varying Times of Digestion (hours) at 37'
Compounds Isolated
Deduced
0
1
4
5
APGP
730
APGP % hydrolysis
730 0%
720 0%
APCP
68
APCP % hydrolysis
445
450
GPCP
256
Gp,G-cyclic-p CP GPCP
0%
0%
478 490 0 485 100%
481 GP/CP
GPCP CPGP
1 150
UPGP
1110
GPUP
285
% hydrolysis
0%
GPCP % hydrolysis
210 0%
CPGP % hydrolysis
1740 0%
1700
UPGP % hydrolysis
350 0%
349
910
925 912 0 918 100%
Gp,G-cycli~-p UP GPUP % hydrolysis
APUP a
70
GP/UP
0% 370 0%
APUP % hydrolysis
211 0
213 0
0% 0%
371 0%
For a description of the procedure see legend to Table V and text.
with the results obtained a t the first stage of digestion of RNA. Thus, the first mono- and small oligonucleotides that appear are Gp and ApGp (stage 1, Table IV). The rate of hydrolysis of several dinucleotides by B . subtilis ribonuclease was also investigated (Table VI). This showed that GpCp and GpUp were split a t enzyme-substrate ratios a t which CpGp, UpGp, ApGp, ApUp, and ApCp were resistant to cleavage. The hydrolysis of GpCp and GpUp is in agreement with the preferential splitting of phosphodiester bonds between guanosine 3'-phosphate and the 5'-hydroxyl terminus of other bases noted above in B. subtilis ribonuclease digests of RNA. As a corollary, the bond between Gp and the neighboring base in Gpterminal oligonucleotides was pwticularly resistant, since a t all stages of digestion (Table IV) oligonucleotides were either free of Gp or contained Gp a t the 3'phosphate end as in ApGp. When a mixture of Gp-terminal (tetra- and larger) oligonucleotides obtained from ribonuclesse-T1 digests was incubated for 2 hours a t 37" with 33,000 units of B. subtilis ribonuclease per mg, no tri- or larger oligonucleotides remained. ApGp, CpGp, UpGp, and some dinucleotides not containing Gp were isolated. Smaller amounts of Ap and Up, but no Gp were found. The absence of free Gp thus confirms the resistance postulated for XpGp sequences (where Xp = Ap, Up, or Cp) in Gp-terminal oligonucleotides noted above. DISCUSSION The B. subtilis ribonuclease described here is comparable to that of Nishimura (1960) with respect to the strain of B. subtilis used as well as to the composition of the growth medium. Our substitution of neopeptone for soybean extract reduced the amount of polysaccharide material but did not noticeably affect
the yield or enzymatic properties of the enzyme. The extracellular B. subtilis ribonuclease is bound on CM-cellulose at acid pH, is stable to acid a t pH 2.0, is not activated by EDTA, and is not significantly affected by CuS04 and MgS04 except a t high concentrations of these metals. Hydrolysis of RNA with B. subtilis ribonuclease involves a cyclic-terminal phosphate intermediate form. At an intermediate stage of digestion (stage 3, Table IV) we obtain a Gp-Ap ratio of 2.9. Furthermore, our B. subtilis ribonuclease has an enzymatic activity (on a weight basis) equal to that of pancreatic ribonuclease. The above results agree with those reported by Nishimura. However, the p H optimum of our enzyme is a t p H 8.5 in Tris-C1 and it is inhibited to about 50% by 0.1 M sodium phosphate p H 8.5. This is ir contrast with the previously reported pH optimum of 7.5 (0.1 M Tris-C1) and lack of phosphate inhibition. In the second paper of this series, devoted to the study of the physical properties of the enzyme (Hartley et al., 1963) the B. subtilis ribonuclease was further purified (2-fold) by exclusion chromatography on Sephadex G-75. Electrophoresis on polyacrylamide a t p H 5 revealed the presence of about 15%of contaminating protein without ribonuclease activity. The p H optimum and enzymatic action on RNA of this more highly purified enzyme did not differ from that of the preparations used in this study, nor from that of peak A or B (Table 11). Since the character of the enzymatic action on RNA remained unchanged but the specific activity was increased a t various stages of the purification process, the presence of two or more enzymes in our B. subtilis ribonuclease may be ruled out. Kickhoefen and Buerger (1962) reported that pancreatic ribonuclease is stable in phenol. With both crude and highly purified B. subtilis ribonuclease, we found that enzyme preparations could be quantita-
Vol. 2, NO.4, July-Aug., 1963 tively desalted and concentrated (10-fold) by phenol extraction (Rushizky et al., 1963). This procedure should reduce the loss of enzyme in the dialysis step (Table 11) and facilitate a further purification of peak A ribonuclease. In our hands, B. subtilis ribonuclease does not show the “simple” purine specificity reported by Nishimura (1960). Instead, the enzyme hydrolyzed -GpGpand -GpAp- linkages in trinucleotides about 100 times faster than any other sequence. If this preference is maintained in RNA, then the enzyme may become a valuable tool (Dekker, 1960) for the isolation and characterization of large oligonucleotides. Given a polynucleotide of chain length 100 and an equal proportion of all four bases, a partial hydrolysis would thus yield (25) (0.5)= 12-13 oligonucleotides which, if present in sufficient quantities, could be isolated and identified with existing methods. Another example of a preferential specificity for neighboring bases (Ap and Tp) was reported for micrococcal nuclease digests of thymus deoxyribonucleic acid (Roberts et al., 1962). The same enzyme preferentially hydrolyzed bonds next to Ap and Up in RNA from tobacco mosaic virus, and was especially effective in attacking those regions of an RNA molecule which contained several adjacent Ap and/or Up residues. Thus, Ap and Up were the fist compounds that could be identified in such digests (Rushizky et al., 1962). The action of the enzyme on nucleic acids appears to be similar to its action on polyadenylic acid (Alexander et al., 1961). In this case, the substrate is first hydrolyzed into large fragments which decreased in sue as the reaction progressed; analogously, the hexamer of pA was preferentially split in the middle giving pApApA and ApApA. There is another application of B. subtilis ribonuclease to the characterization of nucleotide sequences. This involves the resistance of Gp-terminal dinucleotides to hydrolysis. Compounds such as ApGp, CpGp, and UpGp are not split under conditions where other dinucleotides (either free of Gp, or containing Gp but not a t the 3’-phosphate end, e.g., GpCp and GpUp) are hydrolyzed. The base next to Gp in Gpterminal oligonucleotides may thus be identified as shown for the case of a mixture of two pentanucleotides. x ApApCpUpGp y ApApUpCpGp
B . subtilis Ribonuclease x UPGP
CpGp, mono-, and dinucleotides
SPECIFICITY OF
B. subtilis RIBONUCLEASE
793
This procedure is applicable to Gp-terminal compounds, but not to oligonucleotides containing Gp a t a position other than the 3’-phosphate terminus (such as ApGpUp) derived from RNA digests prepared with other enzymes. B. subtilis ribonuclease may thus hydrolyze ApGpUp to Ap and GpUp (Table V). ACJCNOWLEDGMENT We are indebted to Dr. L. A. Heppel for a critical reading of the manuscript and to Dr. N. S. Ikari for help with the growth of B. subtilis. REFERENCES Alexander, M., Heppel, L. A., and Hurwitz, J. (1961),J. Biol. Chem. 236, 3014. Creatfield, A. M., Smith, K. C., and Allen, F. W. (1955), J . Biol. Chem. 216, 185. Dekker, C. A. (1960),Ann. Rev. Biochem. 29, 453. Hartley, R. W., Jr., Rushizky, G. W., Greco, A. E., and Sober, H. A. (1963),Biochemistry 2,794 (paper I1 of this series, this issue). Khorana, H. G., and Vizsolyi, J. P . (1961),J . A m . Chem. Soc. 83,675. Kickhoefen. B.. and Buerger. , M. (1962). . , , Biochim. &oDh.ys. - Acta 6,190.’ Koerner. J. F.. and Sinsheimer. R. L. (1957). . ,. J . Biol. Chem. 228, 1039. ‘ Markham, R., and Smith, J. D . (1952),Biochem. J . 52,558. Nishimura, S. (1960),Biochim. Biophys. Acta 45, 15. Nishimura, S.,and Nomura, M. (1958),Biochim. Biophys. Acta 30, 430. Nishimura, S.,and N o m u a , M. (1959),J . Biochem. (Tokyo) 46, 161. Peterson, E. A., and Rowland, J. (1961),J . Chromatog. 5, 330. Peterson, E.A., and Sober, H. A. (1956),J . A m . Chem. Soc. 78, 751. Peterson, E. A., and Sober, H. A. (1959),Anal. Chem. 31, 857. Roberta, W. K., Dekker, C. A., Rushizky, G. W., and Knight, C. A. (1962),Biochim. Biophys. Acta 55, 664. Rushizky, G. W.,Greco, A. E., Hartley, R. W., Jr., and Sober, H. A. (1963),Biochem. Biophys. Res. Commun. 10, 311. Rushizky, G. W.,and Knight, C. A. (1960),Virology 11: 236. Rushizky, G. W.,Knight, C. A., Roberts, W. K., and Dek-, ker, C. A. (1962),Biochim. Biophys. Acta 55,674. Rushizky, G. W.,Knight, C. A., and Sober, H. A. (1961):, J . Biol. Chem. 236,2732. Rushizky, G. W.,and Sober, H. A. (1962),J . Biol. Chem.. 237, 834. Rushizky, G. W.,and Sober, H. A. (1963),J. Biol. Chem. 238,371. Schmidt, G., Cubiles, R., Zollner, N., Hecht, L., Strickler, N., Seraidarian, K., Seraidarian, M., and Thannhauser, S. J. (1951),J . Biol. Chem. 192,715. Volkin, E., and Cohn, W. E.(1953),J . Bwl. Chem. 205, 767.
